Summarizing batch system stats for the pre GDB from:
https://www.gridpp.ac.uk/wiki/Batch_system_status
It seems that sites split into 3 camps.
Gridengine (5 sites)
* Sites not planning to change.
* Mixture of Univa/son of gridengine and the EPEL version.
HTCondor (3 sites)
* Sites not planning to change.
Torque/Maui (10 sites)
* Lots of problems
* Sites planning to change.
* "Poor Support, Maui intrinsically broken"
* "Maui is unsupported. It had memory leaks. Robert wrote a patch and
there was nowhere to feed it back into. "
* "Maui sometimes fails to see new jobs and so nothing is scheduled "
The Torque/Maui sites are investigating Slurm and HTCondor (one SoGE).
At least 3 sites will follow the consensus.
In detail they are:
Gridengine
==========
UKI-SOUTHGRID-SUSX (Univa)
UKI-LT2-IC-HEP
UKI-LT2-QMUL (EPEL)
UKI-SCOTGRID-ECDF (son of gridengine)
UKI-NORTHGRID-LANCS-HEP Son of Gridengine (HEC),
HTCondor
=========
RAL Tier-1
UKI-SOUTHGRID-RALPP
UKI-SOUTHGRID-BRIS
Torque/Maui
===========
UKI-LT2-Brunel (Slurm and HTCondor in test)
UKI-LT2-RHUL (Will follow the consensus)
UKI-NORTHGRID-LANCS-HEP -> (local cluster - Slurm of HTCondor)
UKI-NORTHGRID-LIV-HEP -> Slurm (Condor?)
UKI-NORTHGRID-MAN-HEP -> slurm
UKI-SCOTGRID-DURHAM -> SLURM
UKI-SCOTGRID-GLASGOW -> HTCondor/SoGE/SLURM
UKI-SOUTHGRID-BHAM-HEP -> Will follow the consensus
UKI-SOUTHGRID-CAM-HEP -> Will follow the consensus
UKI-SOUTHGRID-OX-HEP -> Investigating HTCondor
Now turning to CE support.
CE support
===========
Sites have CREAM. Many investigating ARC.
Multicore
=========
I'm not sure whether it's worth summarizing status on this - but if it's
easy/thought to be worth it, I will.
Chris
|